Optimal Control of an Uncertain Linear Networked Control Systems Under Denial of Service Attacks

Abstract

Controller design based on networked control system (NCS) framework typically relies on the use of computers and communication network systems to automatically monitor and manage the interactions and data exchanges between plant, sensor, controller and actuator elements of the considered closed loop NCS. Such design and implementation scheme are often challenging, particularly due to the requirement of guaranteed reliability and resiliency to possible failure of or malicious cyber attacks on the communication system part. This paper examines the stability of a class of uncertain linear NCS that is subjected to Denial-of-Service (DoS) attacks. In essence, a DoS is a failure phenomenon on the communication links which in NCS framework can prevents the execution of ideal control inputs to take place. Specifically, this paper examines the design of an event-triggered resilient controller for an uncertain linear NCS in the face of such a DoS phenomenon and characterizes sufficient conditions under which the closed loop NCS is guaranteed to remain globally asymptotically stable.